The data myth: interrogating the evidence base for evidence-based peacebuilding

Abstract

This article interrogates three claims made in relation to the use of data in relation to peace. That more data, faster data, and impartial data will lead to better policy and practice outcomes. Taken together, this data myth relies on a lack of curiosity about the provenance of data and the infrastructure that produces it and asserts its legitimacy. Our discussion is concerned with issues of power, inclusion, and exclusion, and particularly how knowledge hierarchies attend to the collection and use of data in relation to conflict-affected contexts. We therefore question the axiomatic nature of these data myth claims and argue that the structure and dynamics of peacebuilding actors perpetuate the myth. We advocate a fuller reflection of the data wave that has overtaken us and echo calls for an ethics of numbers. In other words, this article is concerned with the evidence base for evidence-based peacebuilding. Mindful of the policy implications of our concerns, the article puts forward five tenets of good practice in relation to data and the peacebuilding sector. The concluding discussion further considers the policy implications of the data myth in relation to peace, and particularly, the consequences of casting peace and conflict as technical issues that can be “solved” without recourse to human and political factors.

Policy Significance Statement

The peacebuilding and conflict response sector is large, growing in size, and connected to multiple international and non-governmental organizations. Through the sustainable development goals (SDGs) and the UN’s Triple Nexus of humanitarianism, development, and peace, the sector lies at the heart of much international programming. Much of the sector has enthusiastically embraced the data revolution opening up new possibilities, especially in an era of evidence-based policy. Yet caution is required and this article, after reviewing the drive towards more and faster data, puts forward five tenets of good practice in peacebuilding: disrupting the automatic pilot of data collection; using data for people power; recognizing the limitations of data; maximizing data use; and do no harm.

Introduction

Key stakeholders in the peacebuilding and conflict response sectors emphasize the need for evidence-based policy and programming and often use phrases like “data-driven” or “evidence-based” in their own communications. Data have become paramount for policy and decision-makers to demonstrate scientific evidence that a particular actor or action is having an impact on peace. Without this “evidence-base,” or facts that are supported by a large amount of scientific data, the agendas of conflict response actors can be dismissed as political or not serious, anecdotal, too local, or biased. With this “evidence-base,” conflict response actors seek to be taken seriously as peacebuilders, bringing added value in terms of authoritative data and the legitimacy and funding that comes with that. One example of this is the United Nations (UN) Secretary General António Guterres’ Data Strategy for Action by Everyone, Everywhere with Insight, Impact and Integrity, in which the UN has made data and evidence a priority across its sectors, including in peacebuilding (Jankauskas and Eckhard, 2023).¹ The UN sees its data strategy as delivering “Stronger decision-making and thought-leadership, greater data access and sharing, improved data governance and collaboration, robust data protection, and privacy with respect for human rights, greater efficiency across our work, more transparency and accountability, and more relevant services for people and planet” (United Nations, 2020), but there is little to no recognition of data’s political and agenda setting consequences or the fact that data are often used as a technology of global governance and power (Honig and Weaver, 2019; Davis et al., 2010; Eyben et al., 2015). Crucial to this are the political economies of the peacebuilding sector and the need to compete for market share, produce visualizations and marketing materials, and engage in numerous audit activities—all of which require data.

Whether bilateral or multilateral donors, private donors operating according to the effective altruism agenda (MacAskill, 2017), or civil society organizations seeking to uphold their reputations as peacebuilders, a large industry has grown around the demand for more and faster measurement and evaluation. For this reason, this article interrogates three claims made to the use of data in relation to peace: that more data will lead to better outcomes, that this data must be faster data in order to lead to better outcomes, and that data is impartial and therefore will lead to better outcomes. Taken together, we call this persistent and growing emphasis on more, faster, and neutral data a data myth. The data myth relies on a lack of curiosity about the provenance of data and the infrastructure and power dynamics that produce it and assert its legitimacy.

Therefore, the aim of this article is to question the axiomatic nature of these data myth claims and argue that the structure and dynamics of peacebuilding actors perpetuate the myth. We advocate a fuller reflection of the data wave that has overtaken us and echo calls for an ethics of numbers (Espeland and Stevens, 2008; Saltelli and Di Fiore, 2020; Willis, 2017; Merry, 2016; Littoz-Monnet and Uribe, 2023; Fioramonti, 2014; Best, 2017; Boswell, 2009). In other words, this article is concerned with the evidence base for evidence-based peacebuilding. The transformative potential of data in relation to peace is not in dispute (Carl, 2024; Bell, 2024). Yet the data revolution brings with it a series of practical and ethical concerns that pose questions about peace and power. These include a series of methodological and epistemological assumptions that lie behind prioritizing particular types of data and awarding it an authority or legitimacy that it may not deserve. Littoz-Monnet and Uribe define method regimes as “a special kind of sociomaterial arrangement, which has to do with the production and validation of knowledge” and “directly regulate and control the kind of evidence that is deemed to be accurate and relevant for the governance of global problems” (Littoz-Monnet, 2020, 2). As a result, it is important to unpack the infrastructure and operation of actors functioning in the conflict response space in order to better understand the role of data within it and how the claims surrounding more, faster, and impartial data have become reinforced. We therefore question the utility of data, or information that is known or assumed to be fact and is produced systematically using scientific methods, and processed mostly in the global north, or by elite actors in the global south, for action and uptake (Johnson et al., 2022). We argue that it is important to be inclusive when it comes to knowledge creation around peace because peacebuilding is fundamentally a political and social activity that cannot be completely impartial. Moreover, peace and conflict contexts include highly localized factors that the generalized and aggregated nature of many forms of data tend to flatten out. For this reason, knowledge creation around peace must actively include those experiencing conflict. Data have a role to play in all of this, but it needs to be regarded alongside other sources of knowledge production and subject to a power analysis.

In terms of structure, this article first outlines the processes of datafication that are increasingly shaping the peacebuilding and conflict response sector. It then considers the more data argument, or the belief that more data will lead to better outcomes. It goes on to discuss the faster data argument, or how timescapes represent how time is institutionalized into political and policy processes such as peace processes demonstrating how time is framed, narrated, and imposed to focus on speed and efficiency for better peacebuilding outcomes. It then continues to present a final substantive section about assumptions around data neutrality and impartiality in peacebuilding. Mindful of the policy implications of the foregoing discussion, the article puts forward five tenets of good practice in relation to data and the peacebuilding sector. The concluding discussion further considers the policy implications of the data myth in relation to peace, and particularly the consequences of casting peace and conflict as technical issues that can be “solved” without recourse to human factors. As noted by Wählisch, “personal experience and gut feeling for political; nuances cannot be replaced by machines, yet” (Wählisch, 2020, 123). Essentially, the article is concerned with the risk of data’s over promise. Data have an important role to play in conflict response, but this must be tempered with a consideration of its limitations.

The datafication of the peacebuilding sector

We understand data to be observations that are collected in a systematic manner using scientific methods and that are widely accepted as fact; datafication is built upon the systematic collection, storage, and analysis of that data (Khan, 2023). Such definitions are in keeping with the United Nations Peace and Security Data Hub (https://psdata.un.org/), a product of the Secretary General’s data strategy. The Hub comprises a collection of mainly quantitative datasets using data the UN already collects and stores. Peace-related data usually falls into three categories. The first is situational data that reports on a particular conflict or issue. The second is evaluative data in the form of monitoring and evaluation. The third is related to impact or agenda setting in the form of indexes or barometers, often produced by peace researchers and scholars, such as tracking the implementation of the SDGs. Although these sources of data differ in their provenance and use, they share common challenges in the peacebuilding sector. In this article, we attempt to address all three categories while delineating which category primarily experiences the challenges we highlight.

Peacebuilding is invariably a contested space due to the normative goals implied in the act of building peace. The concept of peacebuilding illustrates a trend towards “mission creep” or an expansion of the theme beyond its original conceptualization. The term was first institutionalized in a 1992 UN Secretary General’s report that referred to “post-conflict peace-building” and saw the concept and practice as “concrete cooperative projects which link two or more countries in a mutually beneficial undertaking” and placed a particular emphasis on demining (Boutros-Ghali and Secretary-General, 1992, 32). Yet peacebuilding has developed far beyond a vehicle for inter-state cooperation, conciliation, and conflict prevention. Indeed, the vast majority of what is often referred to as peacebuilding work takes place at the intra-state level, and peacebuilding programs and projects span a wide range of issues from mental health to primary education and civil society strengthening (Guterres, 2021; Wedge, 2008; Paffenholz and Spurk, 2006). The term remains murky and opaque and its lack of conceptual clarification renders it easily co-optable (Barnett et al., 2007; Paris, 2010).

The challenge for most actors responding to conflicts is to stake a claim as peacebuilders and demonstrate that what they are doing is indeed building peace or at the very least alleviating human suffering and upholding human dignity (Barnett, 2018, 325; Macrae, 2004). In addition, the core of building peace, and what is involved in that endeavor, can often be contradictory among sectors, with the relationship between peace and justice (or the peacebuilding and human rights sectors) often the most contested (Lekha Sriram and Pillay, 2010). This is where data and evidence become crucial and fundamental tools for conflict response actors to negotiate their place within the peacebuilding landscape.

Peacebuilding interventions have always required contextual knowledge, although a number of factors are driving a trend toward scientific data. The first factor relates to neoliberal funding models whereby donors seek value-for-money and verifiable audit trails. This necessitates recipient organizations to collect data using systematic methods so that it can be presented to donors in acceptable (that is, datafied) ways (Coupe et al., 2023). The second, and related, factor is the political economy that peacebuilding organizations occupy. It is a world of competitive tendering and subcontracting that not only necessitates data to demonstrate efficiency and economy but also narratives of success. We have seen the corporatization of the peacebuilding sector, with organizations moving away from initial models of “secular volunteers” (Hoffman, 1998, 13) and “well-meaning amateurs” (Read et al., 2016, 1318) to much more professionalized organizations with corporate structures and processes (Rodriguez-Alarcon and Montoya-Robelo, 2019). Peacebuilding, like much of the third sector, has undergone a “technocratic turn” whereby bureaucratic imperatives play a significant role in shaping organizations (Mac Ginty, 2012). A “managerial rationality” dominates and is empowered to order knowledge in ways that promote and invisibilize different types of data according to format, origin, and message (Bakonyi, 2018, 257).

Also feeding into this datafication is the scale of peacebuilding operations and organizations. Peacebuilding is a substantial sector and its areas of thematic competence have grown significantly. Indeed, the Triple Nexus, or the adoption by the UN, European Union and Organisation for Economic Cooperation and Development of the view that humanitarianism, development, and peace must be addressed in an integrated fashion has led to a broadening of peacebuilding agendas (Howe, 2019; United Nations Trust Fund for Human Security, 2021; Angelini and Brown, 2023; Nguya and Siddiqui, 2020). Peacebuilding has developed from technical assistance after conflict (as envisaged by the 1992 Agenda for Peace document) to become a much more holistic endeavor with multiple thematic interests (Boutros-Ghali, 1992; Firchow, 2020). The key point is that the scale of the endeavor, and indeed the scale of some peacebuilding organizations, awards more opportunities for the deployment of data (Duffield, 2014; Sörensen and Söderbaum, 2012; Howe, 2019; Uvin, 2002; Chandler, 2007; Stern and Öjendal, 2010).

A final point to make in relation to the datafication of the sector is that the digital revolution has provided greater opportunities for the collection, analysis, and comparison of data (Tellidis and Kappler, 2016; Hirblinger et al., 2024; Richmond and Visoka, 2020; Bell, 2024; Firchow et al., 2017). To some extent, this is driven by the supply side or the availability of technology rather than a firm demand for it by potential beneficiaries. We have seen the deployment (often experimental) of remote sensing and programming, artificial intelligence, and visualization technologies in relation to peacebuilding programs (Duursma etal., 2023). While cutting-edge technologies and terms like “PeaceTech” may attract attention, it is worth noting that intermediate technologies—and principally the ubiquity and power of the spreadsheet—still play an enormous role in processes of datafication.

Importantly, the increasing importance of scientific data in the peacebuilding sector reflects a wider story of knowledge hierarchies. Scientific data is a particular type of knowledge that requires not only a technical infrastructure but also a cadre of professionals who can handle and process data, and—crucially—a worldview that takes the systematic nature of data seriously. For Akbari, data is comprised of an “intricate assemblage of organisations, policies, laws, code, software, and platform” (Akbari, 2020, 424). Such assemblages produce and reproduce a cultural landscape that places particular types of data at the top of the hierarchy and, at the same time, downgrades the importance of other types of data—for example, data that may be nonaggregated or considered too localized. Silva points to a “Euroamerican epistemological supremacy” in which scientific, and particularly quantitative, data are valued and regarded as more rigorous than alternatives (Silva, 2019, 94).

Not only are there embedded knowledge hierarchies (Hellmüller et al., 2023) but also these hierarchies serve as something of an epistemological loop. Thus, data helps define the problem, recommend a solution, and thereby reinforce the usefulness of the organization that holds or deploys the data. This has been called “techno-moral power” (Kosmatopoulos, 2021, 258) and draws on Bourdieu’s recognition of how some actors have the power to construct narratives as important or legitimate and others do not (Bourdieu, 1989, 20). Data-powerful organizations are able to set the terms of the debate and establish themselves as gatekeepers over what constitutes the problem, what knowledge is relevant to the problem, and how the problem might be addressed. Daniel Cruz argues that in the peacebuilding sphere, this replicates colonial power in that peace is “only regarded as valid if it represents the values, practices and knowledge that arise from the Global North” (Cruz, 2021, 279).

The increasing emphasis on data in the sector has led to a demand for more and faster data that is treated as impartial. We now turn to demonstrate these three claims made in relation to the use of data in relation to peace.

More data

Data are currency within the peacebuilding sector with the result that many organizations are data producers or commissioners. The growth of the sector, demands for accountability, and the possibilities of technological development all spur demands for more data. The demand also grows because of the need by those organizations to be seen as authorities on peacebuilding or to set agendas for governments and others about how the peacebuilding enterprise should be implemented. Finally, the bureaucratic imperative, much of it driven by donors, has meant that organizations have to collect more data about their own organization and its activities. This might include the collection of equal opportunity statistics on employees or projected spending plans for upcoming financial years. The more data agenda entails more organizations collecting larger amounts of data on a wider range of issues.

The act of measuring is, as already mentioned above, political in that it publicizes the authority and legitimacy of an organization. Its data allows it to assume an authority over a subject, which incentivizes further data collection. For example, Billaud noted of the International Committee of the Red Cross that, “The argument used for turning to statistics was that numbers strengthened the credibility of the ICRC in its dialogue with authorities” (Billaud, 2020, 106). Relatedly, in her research on immigration policy, Boswell demonstrates that data production in organizations serves legitimizing and substantiating functions, where an organization does not only use data to substantiate its claims but also enhances its legitimacy through data collection and production in order to establish “epistemic authority” or the appearance that an institution’s decisions are based on evidence by the mere fact that they are collecting data (Boswell, 2009, 7). In fact, the high demand for evidence-based policy in the sector is led primarily by external, international actors looking to set agendas for those in the global south, in addition to justifying spending to their constituencies (Johnson et al., 2022; Perera, 2017).

In order to procure more funding for efforts in the sector, actors have to justify the reasons why they must continue funding these activities or shift to new ones. More data production and collection help to justify these budgetary decisions and propagate a specific understanding of needs and activities for conflict response, as well as to respond to doubts about aid effectiveness by providing information on the specific contributions of aid to stated policy goals (Eger et al., 2022). Although knowledge about international aid in donor countries is generally limited (Scotto et al., 2017; Wood, 2019), scholars have shown that it is fundamental to present information on aid to taxpayers in order to increase support for aid (Eger et al., 2022; Scotto et al., 2017; Wood, 2019). Experimental studies have proven that information on aid effectiveness is effective in swaying attitudes toward support for aid among donor countries (Eger et al., 2022).

For many years, international agencies like the United Nations Development Program or research institutes like the Uppsala Conflict Data Program or the Stockholm International Peace Research Institute have produced annual and highly respected, statistical updates pertaining to peace, conflict, and development. More recently, however, there has been a significant increase in the number of publicly available metrics. Alongside the already mentioned UN Peace and Security Data Hub which contains over 60 datasets, the peacebuilding data landscape is now populated by a number of global comparative indices such as the Global Peace Index (Institute for Economics and Peace) that claims to present “the most comprehensive data-driven analysis to date”, Freedom in the World Index (Freedom House), P-AX Peace Agreements Database (PeaceRep), the conflict database (ACLED), the European Union Global Engagement Database (Inter-university Consortium for Political and Social Research), the Peace Agreements Database (United Nations PeaceMaker), and the Peace Accords Matrix (University of Notre Dame). Organizations such as SIPRI and the Uppsala Conflict Data Program have augmented their established databases with more specialized ones such as SIPRI’s Multilateral Peace Operations Database or Uppsala’s inclusion of multiple additional variables that can be applied to its main conflict database.

Proponents of data use argue that a boom in the systematic collection of data has helped advance research in the sector. For example, they argue that disaggregated data has given peace and conflict researchers the agency to focus on specific strategies used in armed conflict and that it should be extended to studying peace (Gleditsch et al., 2014; Davenport et al., 2018). However, the links between actual policy changes when it comes to conflict, and the scientific research done on conflict analysis, are tenuous. Data analysis can lead to great recommendations and learning, and many scholars have produced rigorous research with valuable policy recommendations, such as demonstrating the importance of the involvement of women in peace processes (O’Reilly et al., 2015; Paffenholz et al., 2015) or on the relationship between gangs, urban violence, and policy (Jütersonke et al., 2009; Muggah, 2012). However, even when the link between policy and research is clear, research recommendations are often not taken into account, not necessarily for a lack of effort on the part of researchers, but because there are significant political and temporal hurdles, as well as issues of legitimacy and credibility (Millar, 2018a).

The drive for more data occurs in an era in which the United Nation’s SDGs can be regarded as a global strategic plan. The SDGs, including SDG 16 that—in a departure from the Millennium Development Goals—are devoted to peace and justice, are subject to global and country-level tracking. The SDG Tracker, for example, uses “official statistics from the UN and other international organizations” and “allows people worldwide to hold their governments accountable for achieving the agreed goals.” There have also been numerous country-level databases or published metrics that have attempted to track implementation following a peace accord or societal relations following a political transition. Examples include numerous reconciliation barometers (Cole and Firchow, 2019) and numerous efforts at measuring and analyzing peace agreements (Bell and Badanjak, 2019; Joshi and Darby, 2013). Other examples, in post-accord Northern Ireland, include Northern Ireland’s Equality Commission which requires employers to monitor the religious composition of their workforce, equality monitoring for the reformed police force, a peace monitoring report that “measures security, equality, political progress, and cohesion and sharing,” and an Independent Reporting Commission designed to monitor activity by militant and organized crime groups.² The key point is that there is a mounting demand for scientific data to be gathered and published, all requiring resources, expertise, and a specific knowledge infrastructure or methods regime. Much of the data is used or recycled by other peace-interested bodies, suggesting something of a circular economy.

Faster data

This section focuses on the second element of the data myth formulation—that faster data will produce better peace. By faster data, we refer to the concept of speed being applied to all phases of the data cycle: commissioning, planning, collection, collation, analysis, dissemination, and redundancy. The section engages the concept of timescapes that structure the role time plays in the relationship between data and peacebuilding. We argue that some actors are better able to construct timescapes (Holden, 2016, 409) than others; and that timescapes are both constitutive and reproductive of power. Timescapes include temporal features such as time frames, temporality, timing, tempo, duration, sequence, and temporal modalities (past, present, future) and how they relate to space and matter (Adam, 1998, 2008; Meyer-Sahling and Goetz, 2009). A timescape represents how time is institutionalized into political and policy processes such as peace processes, it demonstrates how time is framed, narrated, and imposed by framing the sense of urgency associated with a political process at a given moment. In other words, timescapes refer to the time limits, deadlines, tempo, duration, and so forth of how a conflict response setting is supposed to unfold or a peace agreement should be implemented. Timescapes are produced by power structures, as well as used to produce and enforce those structures and discuss the role that data plays in this relationship.

These power and political relationships and their connection to time encompass what is called chronopolitics, or the role and consequences of time in politics and policymaking (Bergmann, 1992). We are concerned with the temporal politics associated with data and what might be called the fetishization of fast data (Leander and Waever, 2018). While timely data has obvious benefits, especially in fast-moving humanitarian situations, it is worth unpacking issues of power and technocracy that impact the speed with which data can be collected, transmitted, and analyzed. Indeed, the issue of data analysis is crucial, with ever-faster data outstripping the abilities of organizations and researchers to analyze data and integrate it into adaptive planning. As we will see, cultural expectations that equate fast with good outcomes sit in contrast with notions of slow peace or the realization that many conflict response tasks take time and rely on face-to-face interaction and the building of inter-personal relationships (Lederach, 2023). Key to all of this has been a series of technological revolutions that have moved the transmission of data from semaphore flags and dispatch riders to remote sensing, machine learning, and artificial intelligence (Giovanardi and Nicolaïdis, 2022).

Of particular relevance to the temporal aspect of the data myth are cultural understandings of speed as positive and of slowness as negative. Thus, for example, in many societies, a premium is placed on fast service, speedy responses, and early adopters. Conversely, slowness is often associated with backwardness, laziness, and being unbusiness-like (Porritt, 2005, 44). This widespread (but not universal) cultural view of speed as being normatively a good thing has a significant impact on many aspects of life (Tomlinson, 2007, 4). With technology in mind, Eriksen has described “the history of the last 200 years as a history of acceleration” (Eriksen, 2001, 51). For Paul Virilio technology has created a logic that will continuously chase acceleration (Virilio, 2007). Yet, it is important to stress that cultural, individual, and structural traits affect the speed and the rate at which things get done or data is collected. This has consequences not only for the quality of the end product but also for the usage of the data to engage change. In the conflict response field, the trend towards faster data is evidenced by an emphasis on quick-impact projects and short (often one year or less) project and budget cycles. Quick Impact Projects have become an integral part of UN and EU peace programming despite a widespread understanding that peace is a relationship-based process.

Also important to the story of time in relation to peacebuilding and data are tensions between the various timelines that are simultaneously in operation. Peacebuilding is generally thought to be a long-term endeavor that relies on the development of relationships and coming to terms with a violent past and a fragile present. Yet, competing demands on peacebuilding budgets and procurement, accountability requirements, and the limited attention span of donors and the public who fund them, mean that peacebuilding interventions are usually time-limited by the fiscal year, although the actual building of peace can take generations (Firchow and Wingender, 2023).

A key tension in relation to conflict response is the varying—sometimes competing—timelines held by the different actors involved. Indeed, actors from within the same organization or society may have different timelines. Field staff for an aid organization may see a particular issue as requiring urgent attention. They may be supported on this by colleagues in the national headquarters. However, the organization’s chief financial officer, in a far-away capital city, may be operating on a very different timeline. For them, a more urgent issue may be an upcoming financial audit by important funders and the need to make sure that financial probity can be demonstrated. All of the actors will be able to mobilize and present data in support of their position, although some will have more power than others. Key parts of the data myth—the speed at which data is collected and conveyed, the accuracy of the data, and the amount of data—will play a role. Often, the tension between scholarly policy research, which is inherently out of sync with policy timelines, and policymakers needs for evidence is also illustrative of this issue (Hoffman and Weiss, 2011, 266). Important in this regard is the designation of a particular issue as “urgent’.” It is clear that many crises are constructed with some actors having more power than others to frame a particular issue or incident as a “crisis” (Gigliotti, 2020). Again, data will play a key role in the framing of events as a crisis: is the data timely, perceived to be accurate, and does it fit narratives that can be framed as a crisis? Moreover, and specifically in relation to speed, does data come in such quick succession so as to overwhelm the capacity of individuals and organizations to deal with it?

Power dynamics are important here in that much data gathering, and especially the pace of data gathering, is often donor-guided. The demand for evidence-based policymaking is understandable, yet tight deadlines risk compromising the emphasis on high-quality data. In sum, the rhythms of much peacebuilding do not currently coincide with the timing of comprehensive, data-driven processes. This disconnect between donor expectations and the everyday realities of data collection in the field contradicts notions of inclusive approaches to peacebuilding. Donors, power brokers, and governments tend to have the power to set the policy-time clock. This extends to the very local and banal, such as the power dynamic involved in making those with less power wait to signal that their time is less valuable (Auyero, 2012; Mueller-Hirth and Oyola, 2018). Clearly, time in humanitarian settings is not neutral and instead produces specific power relations among actors in complex contexts (Mueller-Hirth and Oyola, 2018, 10).

Impartial data

The essential aim of gathering and processing peacebuilding data is to inform policy and programming around conflict response. The data myth suggests that the more data is produced, and the faster that it is produced, then the better the data and the better the policy and programming. Certainly, there is a logic to this formulation. More data points can produce more accurate data, and timely data may allow for responsive interventions. The way in which data is used and collected within the sector, however, mean that the data landscape is distorted with some data being valued more than others and certain methods of data gathering seen as “more rigorous.” This data hierarchy is very much connected with power and value systems that prioritize “technical knowledge” above knowledge that might be expressed in other formats (Smith, 2008, 238) such as narratives, anecdotes, or stories. This prioritization of technical knowledge rests within a particular culturally constructed knowledge hierarchy because of the widespread use of quantitative measures by North American and European governments. Thus, it is not enough to look at data on its own. Instead, it is important to examine “the social life of data” and the “intricate assemblage of organisations, policies, laws, code, software, and platform” (Akbari, 2020, 413, 424) that enable it. Data, or the particular types of data that are prized by the sector, are not only the products of a particular technocratic culture, but they reinforce it and coproduce it. The sector can be thought of as a complex assemblage of organizations, contexts, personnel, projects, deliverables, and values, all responding to conflict. However, the key to all of this is data or the ability to produce relevant data that can be directed at key targets as necessary.

An ability to produce, analyze, disseminate, and demonstrate a use for data of a particular type signals that an organization in the sector has a specific expertise, one that is valued by donors and peer organizations. It advertises a seriousness and a “right to be at the top tables” while, at the same time, suggesting that data produced in other formats is unserious or does not have the same value. This constitutes an epistemic loop or a one-stop-shop whereby data collected in specific ways helps to define the problem, while the “treatment” of the problem demands that data are collected in particular ways. In so doing, data-capable organizations are reified as legitimate actors in the sector and are thus able to bid for particular calls or are invited to particular forums and are therefore better funded. Despite the diversification of the sector, with many more actors from the Global South, it is still fair to call this a “Western data culture” (Lynch et al., 2023, 914). Such language is not publicly used or acknowledged. In fact, many processes of data extraction are carefully couched in a disinterested language with data processes “constructed as value-less, invisibilized as ‘just sharing’, or naturalized as necessary in order to translate social life into ‘quantifiable’ data of value” (Lynch et al., 2023, 914). Yet behind this neutral language, at least four factors are at work: the casting of conflict response issues as technical, an embedded technocracy, an optimism in relation to technology, and a preference for quantified data. Crucially, although these factors are not obviously political, they suggest particular values and governmentalities. In other words, they are highly political. As Hellmüller and colleagues note, “policies are never just about problem-solving, but also about values and political preferences” (Hellmüller et al., 2023, 1845).

The casting of conflict response as a technical issue that can be “fixed” is ostensibly an act of depoliticization and a belief that change is more likely if the “biased” sector is approached with “unbiased” results about what works for resolving conflicts or improving human rights. Usually, data in relation to peacebuilding is treated similarly to data in relation to the concrete deliverables of humanitarianism or security with expectations of neutral and fact-based findings, such as the number of food parcels delivered or the number of bullets fired, that does not take into consideration the complex, political, and emotional nature of a peace process. In fact, this is an act of politicization dressed as depoliticization and a recasting of processes and events that are caused by structural exclusion, government neglect, and rigged political economies to cauterize curiosity and debate.

The third factor that helps give the impression that data connected to conflict response is without politics has been the enthusiastic adoption of technology by the sector. A techno-optimism, or the belief that technology can solve problems, has been a feature of the sector since its inception. The collection of data at scale and across contexts, as well as the collection of data from hard-to-access contexts, has all been facilitated by technology. The same is true of data processing. Big data requires a minimum level of technological competency and thus a positivity about the digital revolution and the capacity to translate human conditions into data (Richterich, 2018, 28). Behind the apparent value-less nature of “tech-solutionist’ (Lobato and Santos, 2023, 1935) approaches lie political economies of inclusion and exclusion. The application of technological data collection to real-world problems brings with it the promise of clean and objective solutions. Indeed, it holds the promise of democratizing data by possibly “giving a voice to affected people and holding humanitarian organisations accountable” (Fejerskov et al., 2023). Yet the prospect of digital humanitarianism also brings with it the danger of digital discrimination between the digital haves and the digital have-nots.

The fourth factor that helps mask the politics of data has been the emphasis on quantification. Virtually every stage of conflict response involves metrics that can be quantified; needs assessment, planning, logistics, budgeting, monitoring, and evaluation. The advantages of quantified data are that they can convey complex information in accessible ways, allow for simple comparison, enable on-going and end-of-project monitoring, and provide useful visuals. Quantified data can separate complex cases from their contexts to allow for what may be considered objective measurement and assessment. As one proponent of the quantitative study of peace noted, quantitative approaches “would require critically oriented peace scholars to allow for their favorite conjectures and pet ideas about how the world works to be subjected to empirical scrutiny, testing the generalizability of their claims” (Svensson, 2020, 706). Quantified data has a definitive tone. Lanchester notes, “As the House of Commons Treasury Committee said dryly in a 2016 report on the economic debate about EU membership, ‘many of these claims sound factual because they use numbers’” (Lanchester, 2023).

The key point is that much of the data deployed in the conflict response sector is cast as neutral and somehow above politics. This is despite the fact that data are not a stand-alone phenomenon. It is a construct embedded in power relations. Yet data are often unquestioned and awarded authority or status precisely because it is quantified in format and presented as fact.

Five tenets of good data practice in peacebuilding

This discussion of the logic and practice of peace-related data has concentrated on the drive towards more data and faster data that is assumed to be impartial, something we collectively call “the data myth” because of the assumptions that it will lead to better outcomes. We turn to the policy implications of the data myth and how data, and data-commissioning and gathering organizations, might better use data.

Our concerns outlined in this article do not lead us to advocate for a cessation in data production or use, but instead we a caution against the use of data as a panacea for peace. Data serve a purpose when it is produced and used with care and with considerable data literacy. Yet, there is a tendency to emphasize data in the peacebuilding sector even when the utility is not clear. Therefore, in this section, we build on others who have presented guidelines for good data practice for policy (Alliance for Peacebuilding, 2018; Chambers, 2017), and continuing attempts to agree and enforce digital governance (Tauchnitz, 2023). We present our own five tenets of good data practice that we believe are essential to how we can effectively use data for responding to conflicts and building peace. These tenets promote transparency, inclusivity, intentionality, and the proper use of data to contribute positively to peacebuilding efforts. In addition, they advocate for a people-centred approach to data use and production with the goal to give data back to communities in ways that will allow them to use it for their own purposes.

1. Disrupting the automatic pilot of data collection

Data collection is often understood as a necessary element of a peacebuilding project, even if there is no clear goal of how that data will contribute to the outcome. Therefore, our first tenet of good data use for peacebuilding urges us to more intentionally think about how data informs programming and policy in order to reflect seriously about whether or not data collection efforts are actually necessary. We need to regularly question why certain data gets collected, whether or not that data is useful, and whether the effort at collecting that data will make a significant impact on the outcome of the programming or overall impact. When deciding whether data is valuable enough to collect, it is essential to consider the potential impact, costs, and ethical considerations.

In order to determine whether or not data is valuable enough to collect, we have developed questions around fifteen themes, which should be asked before embarking on data collection for conflict response:

Relevance: Is the data directly related to a project or projects? Does it address a specific question or contribute to a clear goal? Does the data collection contribute to fostering peace and positive outcomes?

Purpose: What is the purpose of collecting this data? Who will be informed by this data? How will it be used to inform decisions, solve problems, or advance understanding?

Accuracy: Can the data be collected accurately and reliably, especially in conflict-affected contexts? Are there potential sources of error that could affect its quality? Has the positionality of the researcher or evaluator been considered and any impact it might have on the resulting data and findings?

Cost–Benefit Analysis: What are the costs associated with collecting, storing, and analyzing the data? Do the potential benefits outweigh these costs?

Time: Does the organization have the time to meaningfully analyze the data and transform the data into recommendations? Is it feasible to obtain the data within the desired timeframe?

Availability: Is the data readily available, or will it require significant effort and resources to collect? If not readily available, are participants adequately compensated for their time?

Privacy and Ethics: Does collecting this data raise any ethical concerns or privacy issues? Is it safe for researchers and participants to collect the data? Are there measures in place to protect the rights and confidentiality of individuals involved?

Alternative Sources: Are there existing datasets or sources that could provide similar information? Is collecting new data necessary, or could existing data be repurposed?

Long-Term Value: Will the data continue to hold value beyond the immediate need? Could it be useful for future evaluations, comparisons, or trend analysis?

Data Use and Sharing: How will the data be used, and who will have access to it? Are there plans for sharing the data with other researchers or the public? Can the data be complementary with other data sources?

Data Management: Are there plans in place for organizing, storing, and maintaining the collected data over time? How will the data be managed to ensure its integrity?

Stakeholder Input: Have relevant stakeholders, including those who the data concerns or impacts, been consulted or involved in the decision-making processes necessary to collect and process data?

Unintended Consequences: Are there potential unintended consequences of collecting this data, such as negative impacts on individuals or communities?

Innovation and Learning: Could collecting this data lead to new insights, innovations, or learning opportunities that would not be possible otherwise?

Community Engagement: Is there funding and support for community engagement with the data?

By asking these questions, we can intentionally disrupt the automatic pilot of data collection whereby organizations collect data year on year without asking why the data is necessary. It is important to cultivate a culture of critical thinking and creativity, enabling us to extract meaningful and actionable knowledge from data. A key part of this is a “do no harm” mindset that stops to ask if data, and data collection, are necessary before proceeding.

2. Using data for people power

Often data is used as a tool for accountability to donors, for learning for civil society organizations (CSOs), for academic research, or for public information and funding campaigns. However, data can also be harnessed for communities and everyday people in conflict-affected contexts. Data, when wielded by individuals and communities, have the potential to become a tool of empowerment, enabling them to create consensus, galvanize around an issue, advocate for their rights or needs, amplify their voices, and hold institutions accountable. It shifts the paradigm from everyday people as passive data sources to active agents of transformation. By involving people in data production, collection, and analysis, and finding creative ways to communicate findings, communities have the means to make informed decisions, engage in evidence-based advocacy, and create consensus in order to address pressing social and political issues. Thus, data collection is not merely something imposed on communities, and instead becomes a participative and useful exercise.

Providing data to people in ways that can be useful to them requires creativity and a reorientation of the way we disseminate knowledge based on data collected in the field.

For example, Everyday Peace Indicators (EPI) productively seek to give data back to communities by finding creative ways to actively engage communities with data. They do this through a variety of different forms of engagement, but have found using games or art most effective (Fairey et al., 2023; Fairey et al., 2022). These forms of communication with communities and research participants allow them to not only be recipients of data analysis results but also form dialogs and actively engage with the results to understand how the data may be useful for them. Engaging communities with data to promote people power efforts can be time-consuming and require funding, but it can pay off significantly considering the kinds of grassroots, civil resistance factors that are necessary for creating sustainable political change in conflict contexts (Chenoweth and Stephan, 2011). It also means recognizing the time and energy given by research participants and how it often involves sharing personal information (Field and Johar, 2021).

The Monitoring and Evaluation (M&E) sector has begun incorporating both “learning” and “accountability,” turning M&E into MEL (Monitoring, Evaluation, and Learning), or MEAL (Monitoring, Evaluation, Accountability, and Learning) or DMEL (Design, Monitoring, Evaluation, and Learning) (Urwin et al., 2023). We suggest that there should be another shift in data generation and analysis in the conflict response sector to community engagement. DMEEL (Design, Monitoring, Evaluation, Engagement, and Learning) would require accountability of data use and production to the communities that are served in addition to donors and organizations. It is important that we see DMEEL as a process that can be inclusive, with meaningful community input and engagement. In optimal circumstances, participative data would not simply involve sharing already collected data but would instead involve communities in the commissioning and research design phases of any data-gathering exercise. Otherwise, data collection is just another extractive exercise imposed on communities.

3. Taking positionality seriously

Acknowledging the distinct effects of various forms of identity on knowledge production is a crucial imperative especially in conflict-affected contexts. Identity, encompassing dimensions such as gender, race, ethnicity, socioeconomic status, and cultural background, profoundly influences how individuals engage with the world, interpret information, and contribute to the creation of knowledge. In other words, we must always consider the cultural, social, and political nuances that may impact data collection and analysis. Organizations are often particularly poor at understanding their own positionality and see themselves as neutral. Failing to acknowledge how identity-related factors affect the research process can lead to research (or evaluation) results that may be taken out of context and lack validity. Taking the identity of the researcher and the researched into consideration not only enriches scholarship but also fosters a deeper appreciation of the diverse lenses through which knowledge is crafted, empowering researchers to construct a more equitable and transformative intellectual landscape.

The presentation of data should be transparent about how researcher identity and other factors such as location, timing, and budget can limit the results of research. It is important to provide accessible and comprehensible information about data practices to stakeholders, including donors, local communities, and partners.

4. Maximizing data use

How is data usually used in conflict response? Typically, it is used to inform a conflict analysis or assessment, or to monitor the progress of a project or to inform a report in a final evaluation. Once these tasks are finished, the data usually lose its shelf life and are filed away. However, there are many other ways that data can be reused. With sufficient time spent, data can be transformed into tools that other actors can use for advocacy, dialog, policy, consensus, and so forth For example, the EPIs project has found that different indicators serve purposes for different kinds of actors. In EPI’s Sri Lanka project, the primary purpose was to inform the local programs of a CSO that was contracted by USAID in 20 communities (MacColman et al., 2023). This was done through recommendations presented in a report to the CSO and USAID staff. However, concerns were raised by partners that not all indicators were “actionable” by a CSO or its local partners. This observation encouraged EPI to reflect on the utility of everyday indicators and think more systematically about their various applications, apart from monitoring and evaluation and research. Informed by indicator generation in Sri Lanka in 2022, EPI illustrated the applicability of EPI indicators for different actors and how the data can be used for a variety of purposes beyond monitoring and evaluation (MacColman et al., 2023).

In EPI’s work, the aim is to maximize data usage as a tool for guiding (1) CSO programming, (2) policy, and (3) community engagement. The collection of community-generated indicators through the EPI methodology brings a complementary narrative to standard quantitative data that is based on aggregates and the ironing out of local particularities (Firchow and Mac Ginty, 2017). A key policy implication is to think through modalities of complementarity whereby both qualitative and quantitative data can be given due weight. Multimethod approaches are often stressed in policy and academic research circles, yet in the peace and conflict studies field, there is little conversation or common ground between those who primarily use datasets in peace research, or consider themselves peace scientists, and those more qualitative, mixed-method or practitioner orientated peace researchers (Svensson, 2020; Krause, 2020). Importantly qualitative and participatory data should not be seen as merely illustrative or having the status of a vignette or anecdote that supports a wider “truth” apparently evidenced by quantitative data. Instead, it can be analyzed systematically and indeed quantified. The key point is that participatory and qualitative data should be taken seriously as forms of data and not subordinated in a knowledge hierarchy. The advantage of community-generated data that has been generated through participatory processes is the possibility of local buy-in, not only in data generation but also in activities that catalyze community engagement, helping communities themselves to prioritize issues and take action on them.

Engaging communities around everyday indicators requires dedicated time and organization. This process can be led by local community members or outside actors and can involve creative tools for giving the data back to the people who generated it. In order for communities to fully appropriate their data, local people need time to reflect, discuss, and create plans of action. When this occurs, data have the potential to galvanize community participation and, ultimately, advance local reconciliation processes. EPIs have experimented with various forms of community engagement, for example, using Photovoice (Fairey et al., 2023; Fairey et al., 2022) and card games as a way of giving the data back, encouraging people to reflect on it, and helping to organize action. The main objective of these kinds of activities is to develop and test creative dissemination techniques to enable researchers to share findings with participants and transform the researcher–participant relationship. In doing so, it is possible to move beyond “dissemination” towards “productive engagement,” developing and systematically analyzing tools that make the research process and associated results less extractive and more productive for the populations being researched.³ The distinction between “dissemination” and “productive engagement” is particularly important for communities affected by conflict. Violence and trauma lead to weak social cohesion, tenuous community ties, and low levels of trust in outsiders, heightening the dangers of extractive research processes (Millar, 2018b).

5. Do No harm and conflict sensitivity for data collection and interpretation

Not only can erroneous peacebuilding programming cause harm to communities, but so can data collection and production that does not take into consideration the potential effects it can have on conflict dynamics and efforts at building peace. Therefore, it is important to anticipate potential negative consequences of data collection and analysis and take measures to mitigate them by understanding the limitations of data and how data, and data results, may impact people directly or indirectly.

In the context of data collection, “do no harm” underscores the need to ensure that data collection activities do not put individuals or communities at risk, compromise their safety, or exacerbate conflict dynamics. This principle places a strong emphasis on ethical considerations, participant safety, and the well-being of those directly affected by the conflict. Data collectors must take proactive measures to mitigate risks and prioritize the protection of participants’ rights and dignity. Relatedly, conflict sensitivity involves collecting data in a manner that is cognizant of the potential implications for conflict dynamics, ensuring that data collection does not inadvertently exacerbate tensions or contribute to further divisions.

In essence, “do no harm” is a guiding principle that underscores the ethical responsibility to prevent harm to individuals and communities in conflict-affected contexts. Conflict sensitivity, on the other hand, is a broader framework that encourages a comprehensive understanding of how interventions, including data collection, interact with the complex dynamics of conflict. When it comes to data collection, both principles should inform the design, implementation, and analysis of data collection activities.

Final reflections

This article has argued that the combination of the belief that faster, more, and impartial data will lead to better peacebuilding outcomes can be termed a “data myth.” We refer to it as a myth because we see that there is a need to question the automaticity of the link between more, faster, and impartial data with better outcomes. The article has sought to illustrate the political economies and driving forces behind the data myth, some of which are specific to the peacebuilding and conflict response sectors and some of which are linked with the wider business culture and technological developments that have been adopted by the sectors. Important in all of this is the nature of peace-support programming and the need for it to be contextualized and conflict-sensitive.

The key policy takeaway is data usability. Several hurdles must be overcome to actually use data to implement evidence-based policies in the sector. First, policymakers must be convinced that science is able to improve policy. Achieving this requires education and transparency about the methods and goals behind measurement and evidence. It requires a clear understanding of research and the potential, as well as the limitations, of data on social relations in complex contexts affected by war. It requires data literacy and an understanding of when to use scientific data and when, other, less systematic, data can help to give context and new perspectives.

Second, policymakers must have access to data sources they perceive as legitimate and trustworthy. Consensus is needed on who is a legitimate data source to help improve policy in order to elevate it to an “evidence based” status. In order for better data to lead to better peace, a significant increase in public knowledge production and data analysis needs to be relocated to conflict-affected contexts. Most publicly available data is produced in the global north for research purposes and as such runs into legitimacy problems particularly for local decision-making. The current political economy of data is driven by rich, powerful, and mostly peaceful countries who believe that evidence-based policymaking will make a difference for peace, development, and human rights issues primarily affecting the global south. What is missing, however, are sustained efforts by data-gathering organizations to demonstrate on-the-ground benefits of their data-related activities.

Finally, policymakers must have the resources and time to understand the nuances of data in order to adapt it to their specific contexts. This requires highly trained local researchers with in-depth contextual knowledge to provide assistance in analyzing large amounts of data to make policy recommendations, as well as adapting proven theories of change locally. In order for these shifts to happen, well-trained local researchers of different epistemological and methodological persuasions must be available in order to more readily identify questions the data should address for it to be policy-relevant locally. While the capacity to collect data, with big data efforts in particular, has increased in the last few decades, the capacity to analyze all of that data has not kept up. Although forms of machine learning are assisting researchers to make sense of the massive data we have amassed in the last several decades, this data analysis tool cannot be used for all forms of data analysis and it cannot help us make decisions about the tools we use to analyze the data we produce.

Our aim is not to stymie the use of data and related technologies. Instead, it is to add to calls for governance and responsible use. It is also an exercise in expectation management to acculturate stakeholders to what we can and cannot expect from data. Crucially, knowledge hierarchies, methods regimes, and the power of some data providers mean that there is a danger of inflating the authority of certain types of data, or data methods, and thus regarding it as evidence on which policy is to be based. What is required is a recognition that an evidence base for policy is best served by a scrutiny of data, its sources, and processes. Also important is a recognition of the value of a mixed economy of data whereby multiple types of data (not just top-down scientific data) are taken seriously and is able to inform policy, learning, and accountability, as well as transparently set agendas for peace.